Compare/Kimi K2 0905 vs Grok 4.3 (Non-reasoning)

Kimi K2 0905vsGrok 4.3 (Non-reasoning)

Side-by-side comparison of pricing, 12 benchmarks, and generation speed.

Kimi

Kimi K2 0905

Input
$0.6/M
Output
$2.5/M
Speed
19 tok/s
TTFT
1.83s
xAI

Grok 4.3 (Non-reasoning)

Input
$1.25/M
Output
$2.5/M
Speed
72 tok/s
TTFT
0.57s

Winner by Category

Cheaper
Kimi K2 0905
Faster (tok/s)
Grok 4.3 (Non-reasoning)
Lower Latency
Grok 4.3 (Non-reasoning)
Benchmarks (7-4)
Kimi K2 0905

Pricing Comparison

MetricKimi K2 0905Grok 4.3 (Non-reasoning)
Input ($/M tokens)$0.6$1.25
Output ($/M tokens)$2.5$2.5
Cost for 1M input + 100K output tokens:
Kimi K2 0905$0.85
Grok 4.3 (Non-reasoning)$1.50

Speed Comparison

Output Speed (tokens/s) — higher is better
Kimi K2 0905
19 tok/s
Grok 4.3 (Non-reasoning)
72 tok/s
Time to First Token (seconds) — lower is better
Kimi K2 0905
1.83s
Grok 4.3 (Non-reasoning)
0.57s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
30.931.0
Coding Index
25.925.1
Math Index
57.3
GPQA Diamond
76.7%65.8%
MMLU-Pro
81.9%
LiveCodeBench
61.0%
AIME 2025
57.3%
MATH-500
Humanity's Last Exam
6.3%6.5%
SciCode
30.7%37.4%
IFBench
41.7%47.6%
TerminalBench
23.5%18.9%
Kimi K2 09057 wins
4 winsGrok 4.3 (Non-reasoning)

Frequently Asked Questions

Which is cheaper, Kimi K2 0905 or Grok 4.3 (Non-reasoning)?

Kimi K2 0905 is cheaper overall. Its blended price (3:1 input/output ratio) is $1.07/M tokens vs $1.56/M for Grok 4.3 (Non-reasoning).

Which model performs better on benchmarks?

Kimi K2 0905 wins 7 out of 12 benchmarks compared to 4 for Grok 4.3 (Non-reasoning). See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

Grok 4.3 (Non-reasoning) generates tokens faster at 72 tok/s vs 19 tok/s. However, Grok 4.3 (Non-reasoning) has lower time-to-first-token (0.57s vs 1.83s).

When should I use Kimi K2 0905 vs Grok 4.3 (Non-reasoning)?

Choose based on your priorities: Kimi K2 0905 for lower cost, Kimi K2 0905 for stronger benchmark performance, and Grok 4.3 (Non-reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.